<nettime> On Code and Codework

[Alan Sondheim <sondheim@panix.com>]

On Code and Codework


Consider a well-defined entity x, and its complement -x. Then x^-x = N,
the null set. Consider a second entity y and -y, y^-y = N. Think Nx and
Ny, the null set relativized to x and y. Consider three entities x, y, z,
and take pairs xy, yz, zx. These are symmetrical yx, zy, xz. Let ab stand
for a^b. Then xy, yz, zx are equivalent to null. Let x, y, z divide a
planar region into three regions bordering on each other. Let x@y
represent a line equidistant from the entities x and y. Then x@y, y@z, z@x
all meet at a single point. Divide the plain so that all entities are
grouped in triads; each triad meets in a single point. Divide the plain so
that these single points are grouped in triads and so forth. What branches
are available? Is a single point reached? To operate with x and -x such
that x^-x = N is to operate with discrete entities common to distributive
aristotelian logic. Now consider a second set, X, Y, etc., mapped onto the
first; the mapping may be one-to-one, one-to-many, many-to-one. If the
first, the resulting mapping is reversible. If the second, it is
reversible but the coding itself is not reversible. If the third, the
mapping is not reversible; the result is a set of possibilities, not a
single one. Codes are mappings. There are two types of codes, declarative
and performative. An example of the former is Morse; it is one-to-one, but
all that is produced is equivalence. An example of the second is Perl;
Perl codes procedure. If procedure is coded, then the contents of the
procedure are doubly coded. If a Perl program parses {A} to produce {B}.
then the primary coding is the program which constructs and orders
procedures. The secondary coding is {A} -> {B} which may be considered the
semantic plane of the code. In Eco's A Theory of Semiotics, only a rule
'may properly be called a '_code._' and a rule couples items from one
system with some from another. Eco extends the possibility of code to 'a
set of possible _behavioral responses_ on the part of the destination.
This is performativity. In codework, primary and secondary coding are
entangled. Entanglement may be considered noise in the system. With noise,
the null set N is blurred across fuzzy sets with parasitic inputs; x^-x
and x^y may be and usually are ill-defined. It is this ill-definition -
which functions for example in current definitions of words like 'freedom'
- that tends towards political economy. Political, because culture and the
social are at stake in relation to the definition which is always already
under contestation, and economy, because there are limited resources and
examples for any particular definition. With codework, meaning itself is
problematized as a result of entanglement. In Eco, it is the code which
reaches a destination, not the message. The decoding of the message may or
may not be equivalent to the source. Noise is always already present and
is considered within the channel. This is the T-model of the parasite
described by Serres. Eco states "When a code apportions the elements of a
conveying system to the elements of a conveyed system, the former becomes
the expression of the latter, and the latter becomes the content of the
former. A sign-function arises when an expression is correlated to a
content, both the correlated elements being the functives of such a
correlation.' Code is a collocation or system (not necessarily the same)
of processes; processes are performative; both are temporally-embedded. A
mapping f(x) = y is not temporally-embedded; thus the mapping of the even
numbers onto the number system may be considered an ideality which _is,_
regardless of temporal processes. The structure is given all-at-once
within the formula (and its proof); its proof is a carrying-out of the
truth-value, or a revealing of the truth-value, of the structure. There
are mappings which are systemic, i.e. structure-dependent, and there are
mappings which are non-systemic or purely heuristic, such as randomly
assigning letters in a message to a triple number (page/line/letter-
position) originating from a particular edition of a particular book. In
all of these instances, of course, terms like 'system' and 'assign' are
themselves fuzzy; nevertheless there's a tremendous difference between the
anecdotal and the structural, and there are practical differences in the
ensuing codes and their employment. Peter Gardenfors, in Conceptual Space,
The Geometry of Thought, considers 'conceptual spaces' which are related
to tessellation of the plane. This reminds one of Peirce's simplest mathe-
matics, which is also related to Venn diagrams; in all of these, sets of
entities and concepts in the life-world are mapped into other spaces which
may or may not reflect thinking processes. Geometry is always bound by its
spatial representations; there is no reason to think, at all, that the
mind necessarily works through spatial or any representation for that
matter. Representation is always coded; Sebeok, in Approaches to Animal
Communication, 'Semiotics and Ethology,' points out that the 'model
suggested here entails a communication unit in which a relatively small
amount of energy or matter in an animal (a) the source, brings about a
relatively large redistribution of energy or matter in another animal (or
in another part of the same animal), (b) the destination, and postulates
(c) a channel through which the participants are capable of establishing
and sustaining contact. Maturana somewhere talks about such communication
as the mutual orienting of cognitive domains. Sebeok states that 'Every
source requires a transmitter which serves to reorganize, by a process
called encoding, the messages it produces into a form that can be
understood by the destination. The source and the destination are
therefore said to fully, or at least partially, share (d) a code, which
may be defined as that set of transformation rules whereby messages can be
converted from one representation to another.' As long as one sticks to
transformation rules, code is always procedural. Sebeok states that 'The
string generated by an application of a set of such rules is (e) a
message, which may thus be considered an ordered selection from a
conventional set of signs.' I think that 'ordered' is problematic as well,
since there is clearly a qualitative difference between book-ordering as
described above, and a set of rules based on mathesis. Where does the
arbitrary come in? One might say - and this is an important principle -
that the content of a code itself is directly correlated to its
arbitrariness. In this sense, the measure of a code is related to the
entropy of information within the process of encoding. The greater the
degree of the arbitrary, the more difficult to break, the greater the
entropy and therefore the greater the degree of information within it.
This is not on the level of the double-level of the code, i.e. the content
it operates upon, if there is such content (as in the Perl example above),
but within encoding itself. Every encoding is an encoding of encoding; if
the encoding is fully realized by the product of the code, then its
semantic content / information is low. If Morse encodes a message, more
than likely the message may be decoded based only on the distribution of
letters. The Morse content is low. If a book is used, the decoding is
increasingly difficult and the content of the code is high. This is also
related to issues of redundancy vis-a-vis Shannon and Weaver. Note that a
message and its destination are irrevocably ruptured; there is no
guarantee that an equivalence is attained on any level. Code operates more
often than not on an ontological plane disassociated, or associated by
intention only, with both its source and its decoding; there is no
guarantee that the source and the message-at-its-destination have anything
in common. There is no guarantee of the coherency of the practice of
coding in a particular case, no guarantee of one-to-one or one-to-many or
many-to-one, no guarantee of a zero-parasite-demographics - no guarantee
that the channel, in fact, has not been derailed altogether, as often
happens with bacteriophages. Bateson, in Bateson and Ruesch, Communica-
tion, The Social Matrix of Psychiatry, states 'codification must, in the
nature of the case, be systematic. Whatever objects or events or ideas
internal to the individual represent certain external objects or events,
there must be a systematic relationship between the internal and the
external, otherwise the information would not be useful.' Today one can
say, 'otherwise the information might not be useful,' since it is
precisely in the breakdown of systematic relationships that innovation
emerges.' But meaning may be produced even out of tautology. For example,
propositional logic may be 'derived' from the Sheffer stroke, 'not both A
and B'; it can also be derived from its dual, 'neither A nor B.' What can
we say about these? Only that they represent, as processes or _cullings_
of particular bounded universes, an unbinding/unbounding - 'neither A nor
B' points elsewhere altogether, and 'not both A and B' points either
elsewhere or towards an underpinning of union. At the heart of this
reduction of propositional logic, is a tendency towards dispersion,
towards wandering, the nomadic, even though the symbols within the
calculus proper are completely mute. The Sheffer stroke and its dual, by
the way, are related as well to the processes of inscription with which
this essay began - for what is x^-x, than an _inscription_ of an entity, a
process of coding (and all coding is inscription of one form or another)
the real for the purposes of comprehension, a process that produces, not
only meaning, but _all the meaning there is._ There is no outside to the
sememe, just as there is no landscape without a viewpoint. In this sense
we are bounded, and bound to be bounded. I want to acknowledge and take
responsibility for interpretations here which are necessarily shallow and
possibly misrepresentations as well; this is true in particular of
Gardenfors' book which is complex, and which I have just begun. I have
found the concept of conceptual spaces of use here, as a way of thinking
through code, process, representation, sememe, Eco's planes of expression
and content, etc.; but I do not yet understand it within Gardenfors'
theory. I have also completely neglected what I think is most necessary, a
detailed typology of codes, taking for example temporality into and out of
account in various ways. I cannot see how one can proceed without a deep
reading of Eco's 'Theory of Codes' which is the major section of A Theory
of Semiotics. In the same book, Eco develops a typology of sign production
which is quite useful. Other references might be Barthes' S/Z (although I
constantly find his poeticizing beautiful and problematic), and a quite
useful early book, Symbol Formation, An Organismic-Developmental Approach
to Language and the Expression of Thought, Werner and Kaplan, Wiley, 1963.
Finally, it is clear from all of the above that at best one can sketch a
_discursive field,_ complete with intensifications themselves representing
concepts; this is similar to a loosely-structured Wittgensteinian family
of usages. 'Code,' like 'game,' is always a strategy and a wager from a
theoretical viewpoint, and like much such viewpoints, everything and
nothing is at stake. One would hope for a future of usefulness, politics,
and aesthetics to emerge; the danger, in relation to 'codework' itself, is
that a style develops, and that the uneasy underpinnings - which at least
for me are the most interesting aspects of it - eventually disappear,
absorbed back into issues of genre, etc. Code, like the processes of
postmodernity, is always in a state of renewal, whether or not the 'type'
or 'concept' remains, and at stake within this renewal is our interpreta-
tion of the world itself - our actions and our 'reading' of being and
beings. Wittgenstein's 'silence' at the end of the Tractatus is code's
success, not failure; it is the always already of the always already, but
not its foundation.


===


#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo@bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime@bbs.thing.net